ON THE STRUCTURE OF CERATELLA FUSCA (gRAY). 17 



Undoubtedly in many respects Ceratella and Dehitella call to mind the 

 Hydractiniidfe, but it is doubtful if even our knowledge were confined to that of the 

 structure of the hard parts whether Mr. Carter's classification could be upheld. The 

 one point of resemblance — and at first sight it is the most striking feature — is that 

 the skeleton of both consists of a very irregular branching chitinous network. In the 

 Ilydractiniidfe however this has the form of an encrusting network with at most very 

 feebly developed branches arising from it ; these may more correctly be described as 

 spines and they do not appear to carry any zooids. In the Ceratelladse the whole 

 colony consists of a much-branching structure arising from a comparatively small 

 encrusting root-portion which may itself be made up of branches more or less 

 entwined. In addition to this all the branches bear hydrophores or special 

 developments of the network to support the hydroid zooids. These are never 

 present in the HydractiniidfB but always in the Ceratelladae. Now that the soft parts 

 are known there can be no doubt about separating the two families. The hydroid 

 zooids are quite different those of Ceratella being provided as are those of Coryne 

 with scattered capitate tentacles whilst there is no trace of protective zooids such as 

 are present in Hydractinia and Podocoryne. In addition to this the gonophores 

 arise directly from the ccenosarc and not from modified zooids. The most important 

 points of agreement lie in (1) the existence in Hydractinia and Ceratella of a 

 common external layer of ccenosarc which covers over the whole skeleton mass 

 whether this be encrusting or branching in nature ; (2) the presence in both of a 

 network of ccenosarc tubes forming the hydrophyton. It may however be noted 

 that in both these points we find a similar agreement to exist between Ceratella and, 

 for example, Millepora amongst the Hydrocorallina3 as between the first named and 

 Hydractinia. 



The presence of this external layer which, in the Hydroeorallinae and Ceratella 

 at all events, consists simply of a layer of ectosarc is very difficult to explain. 

 Professor Moseley* has represented it in Millepora as if it formed the outer layer of 

 the surface ccenosarc tubes though even in this case it passes over all the 

 parts (occupied by the calcareous skeleton) which on the surface lie between the 

 tubes, and is very different in appearance and in the relative size of its cells from 

 the ectoderm which clsc^where forms the outer wall of the tubes. In Ceratella it is 

 perfectly independent of the tubes all of which have their own ectoderm covering 

 though at the surface this comes in direct contact with the outer layer. I am not 

 aware of any determination in Hydractinia of the exact relationship of this outer 

 layer though very probably it will he found to agree with that of Ceratella. 



It is not apparently connected in any special way with the formation of the 

 chitinous network as this lies deep within the structure of the branch, and the only 



* Loc. cit., PI. 3., F.gg. 10 and IG. 



